Objectives This paper aims to study the influencing factors and mechanisms of ship model resistance prediction uncertainty, as well as proposing corresponding suppression methods.

Methods Resistance tests and local flow measurements on the flow around KCS models of 1.725 m and 3.450 m are carried out in a circulating water channel. The effects of the model scale on the resistance predictions are obtained by comparing the results of the two tests. Numerical simulations of the flow around the 3.450 m model in an unbounded water area and circulating water channel are achieved. The results under the two conditions are compared to analyze the blockage effect and validate the simplified Tamura correction formula. Numerical simulations of the flow around the 3.450 m model under different turbulence intensities in a circulating water channel are then carried out to analyze the effects of turbulence intensities on resistance predictions.

Results The results show that the resistance prediction uncertainty of small ship models due to statistics collections and inflow uniformity can be amplified if the inflow velocity is too low. This demonstrates that the blockage effect of the large model could increase hull sinkage, pressure gradients on the hull and wave amplitudes, leading to an increase in ship resistance. The average difference between the resistances predicted in the unbounded water area and circulating water channel is 4.56% without correcting for the blockage effect. The difference is reduced to 2.25% with the correction of the simplified Tamura formula. The total resistance increases by an average of 3.75% as the turbulence intensity increases from 1% to 2%. It is observed that the decay of turbulence intensity along the flow direction is linear from the entrance to the bow.

Conclusions This study shows that reasonably large models should be used wherever possible and the turbulence intensity should be strictly controlled to reduce the uncertainty of ship resistance tests in circulating water channels. In addition, the simplified Tamura formula should be used to correct the blockage effect, and the appropriate inflow turbulence intensity should be used in numerical prediction according to the dissipation of the turbulence intensity.